The ABINIT Users Mailing List ( CLOSED )

["Bart Partoens" <bart.partoens@ua.ac.be>]

Hello Sergey,

In the follow up to your question it was suggested to add "prteig 1".

In my opinion it is due to a small typo in the file "prteigrs.F90"  in 
version 4.4
(and maybe also 4.5 but I did not check it) of Abinit. This file is in the 
directory
"abinit/Src5_common" .

In the description of the keyword "prtvol"
http://www.abinit.org/Infos_v4.4/varfil.html#prtvol
it is written that, when you do a non-self-consistent calculation "iscf -2", 
the information
for all k-points will be written out, even when you choose prtvol 0 or 1 . I 
noticed that
this is new from version 4.4.

However, in the routine "prteigrs.F90" it is written (line 109 and 110 in 
v4.4.4)
  nkpt_eff=nkpt
  if( (prtvol==0.or.prtvol==1.or.iscf==-2) .and. 
nkpt_eff>nkpt_max)nkpt_eff=nkpt_max
Thus when iscf equals -2, the maximum number of k-points for which 
information will
be written is nkpt_max ( = 50). Sergey, I see that in your input file 
zb_band.in
you do indeed a calculation for "iscf -2".

So this may be the typo, the problem is solved if the line 110 is changed 
into
  if( (prtvol==0.or.prtvol==1.or.iscf/=-2) .and. 
nkpt_eff>nkpt_max)nkpt_eff=nkpt_max

In version 4.3 there is no problem, because in file "prteigrs.f", the 
corresponding line is
if( (prtvol==0.or.prtvol==1) .and. nkpt_eff>nkpt_max)nkpt_eff=nkpt_max
So the feature with "isc -2" is not included yet in v4.3.

Regards,

Bart Partoens


----- Original Message ----- 
From: "Sergey Lisenkov" <proffess@yandex.ru>
To: <forum@abinit.org>
Sent: Saturday, May 14, 2005 2:44 PM
Subject: [abinit-forum] All k-points are not printed when use prtvol=10


> Dear Abinit users and authors,
>
> I am a really confused when I use 'prtvol 10' option for printing all 
> k-points during band structure calculations. The total number of k-points 
> is 110. But in output file I see only 50 k-points in Dataset 2. My input, 
> output and log files are attached. The version of abinit is 4.4.4 
> (parallel version).
>
> Many thanks,
> Best wishes,
>   Sergey